Spinal Dynamics I: The Axio-atlanto-occipital
Assemblage
Bones interact through joints. The relative placements of
bones across joints determine how they move in space. In this
section we will consider two related assemblages of bones. The
first is the atlanto-occipital and atlanto-axial joints, which
together marry the skull to the vertebral spine.
The movements in the axio-atlanto-occipital assemblage are
complex because of the interactions between these bony elements
(Kapandji 1974; Hoppenfeld 1976; White and Panjabi 1978; Dvorak
and Panjabi 1987; Dvorak, Penning et al. 1988; Dvorak, Schneider
et al. 1988; Nordin and Frankel 1989; Panjabi and White 1989;
Penning 1989; White and Panjabi 1990; Ghanayem, Zdeblick et al.
1998; Panjabi, Dvorak et al. 1998; Bogduk 1999; Bogduk and
Mercer 2000; Levangie and Norkin 2001; Standring 2005; Langer
2005m; Langer 2005p). The movements in this joint complex are
unlike those in the rest of the spine, both in type and
magnitude of angular excursion.
In related matters, the vertebral artery passes between the
transverse processes of the atlas and the axis, which places it
far from the center of rotation for those two bones, and the
amount of rotation between those bones in much greater than
between any other vertebrae. Both of these anatomical factors
place the segment of the vertebral artery between the two bones
at substantial risk of excessive strain, making it the part of
the artery most apt to be injured by physical activity or
therapeutic manipulation. We will briefly consider this anatomy
and its implications for blood flow as part if our consideration
of the axio-atlanto-occipital assemblage. Blood flow is
considered in considerably greater detail elsewhere (Langer 2005m;
Langer 2005n; Langer 2005o; Langer 2005p).
The second bony assemblage is the remainder of the cervical
spine, the lower cervical spine. The lower cervical spine is
one of the more flexible sections of the spinal column and it
moves about an interesting type of facet joint that allows
movements that do not fit easily into any of the cardinal planes
(Kapandji 1974; White and Panjabi 1978; Nordin and Frankel 1989;
Panjabi and White 1989; Penning 1989; White and Panjabi 1990;
Milne 1991; Committee 1998; Ghanayem, Zdeblick et al. 1998;
Onan, Heggeness et al. 1998; Panjabi, Dvorak et al. 1998; Bogduk
1999; Bogduk and Mercer 2000; Levangie and Norkin 2001;
Standring 2005; Langer 2005q; Langer 2005r; Langer 2005s). The
facet joints of the second cervical vertebra,C2, to the first
thoracic vertebra, T1, are inclined at about a 45° angle to the
coronal plane of the vertebra. As a result, the movements of
1
the cervical spine involve a combination of sideflexion and
lateral rotation. Also, since the successive vertebrae are
inclined relative to each other, this movement has different
consequences for the movements of the cervical spine at multiple
levels. The lower cervical spine provides an excellent
opportunity to model an assemblage of bones and joints, because
it is essentially the same element repeated several times, which
greatly reduces the numbers of parameters that need to be
manipulated.
The Axio-atlanto-occipital Assemblage
Anatomy: The inferior aspect of the second cervical vertebra,
the axis, is like the other cervical vertebrae in having
inferior facets that face anteriorly, inclined about 45° to the
coronal plane and an intervertebral disc that joins two
vertebrae in an uncinate (U-shaped) joint. The upper surface of
the axis is quite different. We will deal with that type of
joint in some detail, below.
The rest of the bone is of more
interest at this point. The axis has a long, tooth-like,
odontoid process, or dens, that extends rostrally, through the
ring-like anterior arch of the atlas, the first cervical
vertebra. The odontoid process is the embryological remnant of
the vertebral body of the atlas, which has become fused to the
vertebral body of the axis. It is held in place by a transverse
ligament that passes posterior to it, joining the two sides of
the atlas and completing a tubular socket that encloses the peglike odontoid process. The axis and atlas vertebrae abut upon
each other in a pair of lateral facet joints that form the
lateral wall of the tube and allow and support lateral rotation
between the two vertebrae. The lateral facets are tilted
posteriorly and laterally, so that the support surface is
slightly inclined.
The odontoid process provides a fulcrum for lateral rotation
of the atlas upon the axis. Consequently, the axis of rotation
for the atlanto-axial joint lies within the odontoid process.
During lateral rotation, the lateral facets slide upon each
other. The surfaces of the facets in this joint are almost
flat, only slightly convex in their facing surfaces. The modest
curvature of both facet surfaces allows the two bones move a bit
closer as they side over each other, allowing a small amount of
additional rotation. The overall movement between the two bones
is rather like stacked funnels, where the upper funnel rotates
about the spout of the lower funnel.
2
There is also a small amount of play in the odontoid part of
the atlanto-axial joint so that the atlas may tilt forward and
back through about 10° of sagittal movement. That is a
comparatively small amount of movement when compared to the full
range of sagittal movement is the cervical spine, but comparable
to amount of rotation in most individual intervertebral joints.
The principal movement of the atlas about the odontoid process
of the axis is lateral rotation. It is estimated that the
average range of movement is a bit more than 40°, varying
between about 30° and 55° in different individuals, which means
about 45° in either direction, for a total of about 90° of
lateral rotation in the joint. That is a remarkable amount of
movement in a single vertebral joint and it is physiologically
important in that the vertebral artery passes just lateral to
the lateral articular processes, so it is stretched between the
transverse foraminae of the two vertebrae. Consequently, the
most common site for mechanical trauma to the vertebral artery
is in the segment that passes between C2 and C1.
The joint between the atlas and the occiput, the atlantooccipital joint, occurs between the lateral articular processes
of the atlas and the occipital condyles. The surfaces of those
facets are convex inferiorly, so that the occipital condyles
3
effectively sit in a shallow bowl formed by the lateral facets
of the atlas. The axis of rotation is transverse and located
slightly rostral to the joint, passing through the occiput. The
two joint surfaces converge anteriorly, which causes the joint
to become tighter as the head tilts into flexion or extension.
The range of motion for sagittal movements is estimated to be
between 10° and 30°, with the average about 20° of angular
excursion. There may a small amount of sideflexion, about 3°,
and lateral rotation, about 5.5°, however these may be more
joint play than a useful physiological movement. However, the
small amount of sideflexion turns out to have physiological
implications for the amount of sagittal rotation in the atlantoaxial joint (see below).
The rings are, from above down, the occiput, the
atlas and the axis. The lateral processes of the
vertebrae are indicated by the pinched regions of their
rings. The occipital ring is at the level of the axis
of rotation and the pinched region is the location of
that axis. All the bones are in neutral position. The
r axis is anterior/posterior and the s axis is
medial/lateral. The scale is arbitrary, related to the
physical dimensions of the atlas.
Modeling the Anatomy with Quaternions: Given these anatomical
characteristics we can express the anatomy of these joints as a
set of three framed vectors, one for each bone, and a set of
rotation quaternions with definite locations that describe the
axes of rotation for the various joints. Since the axes of
rotation travel with the bones, they are effectively elements of
4
the framed vectors. Such a model is given in some detail
elsewhere (Langer 2005m).
In the following discussion, based on that model, the atlas
and axis vertebrae and the occiput have been represented by
simplified images, pinched tori, which capture their locations
and orientations without obscuring their spatial relationships.
Given the rotation in each direction for each element of the
joint, the model computes the arrangement of the bones.
In neutral position, the three bones are aligned with the
lateral processes of the cervical vertebrae in the same coronal
plane and the transverse foramina of the atlas lateral to that
for the axis. The pinched regions of the rings occur at the
locations of the transverse foraminae. The anterior aspects of
the vertebrae are to the right. The occipital ring is placed so
that the axis of rotation is through the pinched parts of the
ring and the ring is horizontal in neutral position. The ring
would sit some distance into the skull, rather than at the level
of the occipital condyles.
In the following figure, rotations have occurred in several
of the possible axes of rotation. Rotations occur both within
the assemblage and outside it. The lower neck is assumed to
have caused the axis to be tilted 45° anteriorly and laterally
rotated 45° to the left. The axis is laterally rotated 20° to
the left and the occiput is anteriorly tilted 20°. Any
combination of rotations may be examined in the model, making it
a tool to study the movements in the upper cervical spine.
5
The placements of the occiput, atlas, and axis after
the indicated movements about several axes of rotation
The Effect of Movements Upon the Intracranial Segment of the
Vertebral Artery: Once one has a model of this sort, it is
possible to perform calculations that address anatomical
questions. For instance, the vertebral artery passes behind the
facet joints for the atlanto-occipital joint and enters the
vertebral canal to run rostrally through the foramina magnum to
the pons in the brainstem, where it fuses with the opposite
vertebral artery, to form the basilar artery. The model allows
one to determine how flexion and extension of the occiput upon
the atlas affects the length of the vertebral artery required to
bridge that gap. An extension vector attached, to the occiput,
was created to encode the location of the pontomedullary
junction within the skull. Another extension vector, attached
to the atlas, encodes the location of the vertebral artery as it
passes behind the articular facets and is apparently firmly
attached to the bone. By examining the distance between those
two locations as flexion and extension in the atlanto-occipital
joint are varied, one can obtain a fair estimate of the length
of the vertebral artery that is necessary to accommodate for the
movement.
It turns out that the placement of the artery adjacent to the
facet joint minimizes the change in distance between the end
points of the artery. It is estimated that the range of motion
is less than 20° for both directions combined, so, the change in
6
the gap between the extremes of the physiological sagittal
movements is about 10% of the gap in neutral alignment. Given
that the artery is not normally taut, that is a comparatively
small change in distance and we would not expect there to be
significant strain in that segment of the artery.
The gap between the lateral facets of the atlas and
the fusion of the vertebral arteries has been computed
for a range of flexion and extension. The shaded
interval corresponds to actual physiological values.
The Effects of Movements Between C1 and C2 upon the Vertebral
Artery: We can ask a similar question about the distance
between the transverse processes of the first and second
vertebrae. The question is addressed by creating extension
vectors for the transverse foraminae of the two vertebrae.
Since the amount of rotations between the vertebrae is
considerable, about 45° in each direction, and the artery is
located far from the axis of rotation, there is considerable
potential for strain in the artery.
7
The red line segments indicate the locations of the
vertebral arteries between the atlas and the axis when
there is a rotation between the two bones.
When the calculations are performed it can be seen that there
is a substantial increase in the distance that the vertebral
artery must span as the rotation occurs. As shown in the
following figure, the distance between the transverse process
increases about 1.35 to 1.5 times as the neck approaches full
physiological lateral rotation. That is more than the artery is
able to stretch without tearing. However, arteriograms of the
arteries in many persons indicate that there is normally
substantial laxity in the artery when the neck is in neutral
position, so that the rotation simply takes up the slack instead
of straining the artery.
Still, the most common site for traumatic damage to the
vertebral artery is in the interval between the atlas and the
axis. Should the neck be rotated more than it normally does,
there is the potential to strain the artery to the point where
it is torn. That may become a consideration when a manipulation
of the neck is done that suddenly increases the range of motion
in the atlanto-axial joint, if the length of the artery had
adjusted to a shortened maximal gap.
8
Rotating the head to the endrange of lateral rotation is
sufficient in some individuals to cause a cessation of blood
flow in a vertebral artery (Arnold, Bourassa et al. 2003; Arnold,
Bourassa et al. 2005; Arnold, Bourassa et al. 2005; Langer 2005n; Langer
2005o; Langer 2005p). If that is combined with sideflexion in the
atlanto-occipital joint, then nearly half of the normal
individuals tested show a significant reduction or cessation of
blood flow. Although the sideflexion is a small movement it
increases the amount of rotation in the atlanto-axial joint.
That maneuver is frequently, used to increase the movement in
the atlanto-axial joint to endrange when manipulating the neck.
The length of the gap spanned by the vertebral
artery is plotted versus the amount of rotation in the
atlanto-axial joint. Normal range is about 40° in both
directions, but it may be as much as 60° in some
individuals.
The Role of the Alar Ligaments in Movements of the Upper
Cervical Spine: Another situation where the gap may increase
substantially is when the lateral rotation is increased by
9
changing the usual physiological constraints upon the movement
(Langer 2005p). Experimental studies indicate that if the
atlanto-occipital joint is sideflexed before or concurrently
with lateral rotation of the atlanto-axial joint, then there is
a significantly increased rotation in the atlanto-axial joint.
It is suspected that it is the relaxation the alar ligament on
the sideflexed side that is responsible for the increased range,
because the alar ligaments have the potential to restrict the
movements of the odontoid process in the atlanto-axial joint.
That is also a question that may be addressed in this model of
the upper cervical assemblage.
The atlas and axis are joined to the base of the occiput by
several fascial membranes and ligaments, which enclose the
brainstem and upper spinal cord and lend stability to the
region. However, one of these, the alar ligaments, is thought
to play a part in restricting the amount of lateral rotation
between the axis and the skull. The alar ligaments extend from
the inner rim of the foramen magnum, just medial to the
occipital condyles, to the apex of the odontoid process of the
axis vertebra. Because the atlas lies between those two bones,
the alar ligament must affect it and we expect to see
consequences both for lateral rotation, which occurs largely
between the atlas and the axis, and for sagittal head movements,
which occur largely between the atlas and the occiput.
The alar ligaments bind the occiput to the odontoid
process of the axis. Consequently it is able to
influence flexion/extension and lateral rotations.
The alar ligaments are longer than the gap between its bony
attachments, therefore it is possible to move the attachment
sites for some distance before one or both ligaments become taut
and stop the movement. Flexion and extension move the occipital
insertion anterior and posterior relative to the odontoid
process and therefore one would expect their range to be
restricted by the tethering that occurs when the gap between the
two insertions becomes equal to the length of the ligament.
10
Similarly, lateral rotation between the occiput and the axis
will move the attachment on the odontoid process away from the
attachment on the occiput.
Since the center of rotation for the atlas upon the
axis is through the odontoid process, a lateral
rotation will move the occipital insertion away from
the odontoid insertion on the side opposite the
direction of the rotation and move the two attachments
closer together on the opposite side. Consequently,
the alar ligaments are checks on lateral rotation.
11
When there is no sideflexion lateral rotations are
symmetrical and the alar ligaments appear to be the
principal restraint on rotation. With small amounts of
sideflexion, ipsilateral rotation is substantially
reduced but the alar ligaments cease to be a restraint
on contralateral rotation. The horizontal gray line
indicated the usual length of an alar ligament.
Since we know the relative distances between significant
structures in the upper cervical assemblage and the approximate
length of the alar ligaments, it is possible to enter the values
into the model of the upper cervical assemblage and to examine
the implications for movements between the bones in the
assemblage. That has been done and the details are available
elsewhere (Langer 2005p). We will examine the results of one set
of calculations in which the distance between the insertion
sites for the alar ligaments are computed for a range of lateral
rotation, assuming that the head started in neutral position
(the blue lines), sideflexed 3° (green lines), or sideflexed 5°
(red lines). The gap for each ligament is computed. When there
is no sideflexion, the relationships are symmetrical (blue
lines) and the head can rotate about 40-45° before it is stopped
by a taut alar ligament. That is about the range of lateral
rotation that is normally observed. If the head is sideflexed
(green and red lines), then the relationship is asymmetrical and
shifted so that there is less ipsilateral rotation before the
ligament becomes taut, but more contralateral rotation. In fact
the alar ligament effectively ceases to be a restriction upon
lateral rotation to the contralateral side.
Other ligaments must come into play, because at a point not
that much greater than 45° of lateral rotation the lateral
facets will fail to abut and there is a possibility of them
subluxating and locking in a side-by-side position. There is
also a risk of impingement of the vertebral arch upon the
cervical spinal cord. Occasionally, individuals tear one alar
ligament and it is observed that they gain a substantial amount
of lateral rotation, but the rotation is stopped short of
catastrophic impingement, although their facet joints can become
locked in endrange lateral rotation.
The model of the upper cervical assemblage predicts the
approximate range of lateral rotation without sideflexion and it
indicates that the increased range when the head is sideflexed
may well be due to the reduction of the gap that occurs with
sideflexion. Not considered here is the observation that the
model predicts that when the head is flexed or extended in the
atlanto-occipital joint, then there is a substantial reduction
in the amount of lateral rotation that can occur before the alar
ligaments become taut and stop the movement (Langer 2005p).
12
Arnold, C., R. Bourassa, et al. (2003). "Doppler studies
evaluating the effect of a physical therapy screening
protocol on vertebral artery blood flow." Manual Therapy
9/1: 13-21.
Arnold, C., R. Bourassa, et al. (2005). Changes in Blood Flow Associated With
Vertebral Artery Stress Tests.
Arnold, C., R. Bourassa, et al. (2005). Dependence of Vertebral and Carotid Artery
Flow Upon Vertebral Artery Stress Tests and Premanipulative Holds.
Bogduk, N. (1999). "The neck." Baillieres Best Pract Res Clin
Rheumatol 13(2): 261-285.
Bogduk, N. and S. Mercer (2000). "Biomechanics of the cervical
spine. I: Normal kinematics." Clin Biomech (Bristol, Avon)
15(9): 633-48.
Committee, T. C. S. R. S. E. (1998). The Cervical Spine.
Phiadelphia, Lippincott - Raven.
Dvorak, J. and M. M. Panjabi (1987). "Functional anatomy of the
alar ligaments." Spine 12: 183-190.
Dvorak, J., L. Penning, et al. (1988). "Functional diagnostics
of the cervical spine using computer tomography."
Neuroradiology 30: 132-137.
Dvorak, J., E. Schneider, et al. (1988). "Biomechanics of the
craniocervical region: the alar and transverse ligaments."
J Orthop Res 6: 452-461.
Ghanayem, A. J., T. A. Zdeblick, et al. (1998). Functional
Anatomy of Joints, Ligaments, and Discs. The Cervical
Spine. T. C. S. R. S. E. Committee. Philidelphia,
Lippincott - Raven: 45 - 52.
Hoppenfeld, S. (1976). Physical Examination of the Spine and
Extremities. New York, Appleton-Century-Crofts.
Kapandji, I. A. (1974). The Physiology of the Joints. Annotated
diagrams of the mechanics of the human joints. New York,
Churchill Livingstone.
Langer, T. P. (2005m). Movements of the Upper Cervical Assembly: A Model of the Axioatlanto-occipital Assembly. :
http://homepage.mac.com/tlanger_sasktel_net/CervicalSpine/u
ppercervical.html.
13
Langer, T. P. (2005n). Strains in the Vertebral Arteries With Movements in the Upper
Cervical Spine. :
http://homepage.mac.com/tlanger_sasktel_net/CervicalSpine/u
ppercervical.html.
Langer, T. P. (2005o). Blood Flow in the Vertebral Arteries When Strained by Lateral
Rotation in the Upper Cervical Spine. :
http://homepage.mac.com/tlanger_sasktel_net/CervicalSpine/u
ppercervical.html.
Langer, T. P. (2005p). Why does sideflexion increase ipsilateral vertebral artery occlusion
with contralateral atlanto-axial rotation? :
http://homepage.mac.com/tlanger_sasktel_net/CervicalSpine/u
ppercervical.html.
Langer, T. P. (2005q). Building a Neck and Watching It Move: Anatomical Foundations
for the Description of Movements in the Cervical Spine.:
http://homepage.mac.com/tlanger_sasktel_net/CervicalSpine/l
owercervical.html.
Langer, T. P. (2005r). A Model for the Study of Movements of the Lower Cervical Spine. :
http://homepage.mac.com/tlanger_sasktel_net/CervicalSpine/l
owercervical.html.
Langer, T. P. (2005s). Movements in Artificial Necks.:
http://homepage.mac.com/tlanger_sasktel_net/CervicalSpine/l
owercervical.html.
Levangie, P. K. and C. C. Norkin (2001). Joint Structure and
Function. A Comphrehensive Analysis. Philadelphia, F. A.
Davis Company.
Milne, N. (1991). "The role of zagapophyseal joint orientation
and uncinate processes in controlling motion in the
cervical spine." J. Anat. 178: 189.
Nordin, M. and V. H. Frankel (1989). Basic Biomechanics of the
Musculoskeletal System. Philadelphia, Lea & Febiger.
Onan, O. A., M. H. Heggeness, et al. (1998). "A motion analysis
of the cervical facet joint." Spine 23(4): 430-9.
Panjabi, M. M., J. Dvorak, et al. (1998). Cervical Spine
Kinematics and Clinical Instability. The Cervical Spine. T.
C. S. R. S. E. Committee. Philadelpia, Lippincott - Raven:
53 - 78.
14
Panjabi, M. M. and A. A. White (1989). Biomechanics of Nonacute
Cervical Spinal Cord Injuries. The Cervical Spine. T. C. S.
R. S. E. Committee. Philadelphia, J.B Lippincott: 91 - 96.
Penning, L. (1989). Functional Anatomy of Joints and Discs. The
Cervical Spine. T. C. S. R. S. E. Committee. Philadelphia,
J.B. Lippincott: 33 - 56.
Standring, S. (2005). Gray's Anatomy. The Anatomical Basis of
Clinical Practice. Edinburgh, Elsevier, Churchill
Livingstone.
White, A. A. and M. M. Panjabi (1978). Clinical Biomechanics of
the Spine. Philadelphia, J. B. Lippincott.
White, A. A. and M. M. Panjabi (1990). Clinical Biomechanics of
the Spine. Philadelphia, J.B. Lippincott.
15